127,664Total entries

60Added by human

127,604Found by agent

13Categories

Knowledge catalogue

Search: “models”

95,172 results

Safety

Revisiting scaling laws for reward optimization

DGX agent

arXiv:2609.38526v1 Announce Type: cross Abstract: Scaling laws for optimization against reward models in AI alignment have pinned down how performance depends on optimization effort---measured by a KL

safetyarxiv-cs-ai

1 Oct 2026

Model Releases

Scaling Laws for Looped Mixture of Experts

DGX agent

arXiv:2609.40316v1 Announce Type: cross Abstract: Looped transformers and Mixture-of-Experts (MoE) offer complementary routes to efficient scaling: recurrence increases computational depth at fixed pa

Search: “models”

Revisiting scaling laws for reward optimization

Scaling Laws for Looped Mixture of Experts

SceneJail: Exploiting Video Scenario Context to Jailbreak Multimodal LLMs

Sense and Sensitivity: Benchmarking LLM Clinical Triage Recommendations with Physician Experts

UniST-Pred: A Robust Unified Framework for Spatio-Temporal Traffic Forecasting in Transportation Networks Under Disruptions

When Scientific Contradictions Are Lost in Translation

WorkGenesis: Building the Worlds That Teach Agents to Work

AdviSD: Learning to Advise Frontier LLMs via Targeted Multi-Turn Self-Distillation

BITEM at the NTCIR-19 R2C2 Task: Predicting Confidence from Agentic RAG Pipeline Signals

Breakthroughs on every frontier. I can't wait to share more on what Praxis-1 will unlock across industries.

Counterfactual Probing for Parallel Unmasking with Hidden Forest Structure

CounterSteer: Suppressing Indirect Prompt Injection with Activation Steering

Cropland PAtteRNS: Parallel Dimensional Attention Networks and Attention to Dataset Disparity for Crop Segmentation in Satellite Imagery Time Series Data

CruxBench: A Benchmark of Information Discovery

Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution

Does the Unsafe Gradient Survive a Conversation? On the Fragility of Gradient-Based Jailbreak Detection in Multi-Turn Dialogue

Evaluation Choices Shape Biomedical ML Claims: A Pediatric Pneumonia Benchmark Case Study

FedLAFP: Low-Rank Aggregation Meets Full-Rank Personalization in Federated Fine-Tuning

Follow-up: my native Rust + Vulkan Transformer training backend — 14 days later, now 14 parity-verified architectures and full PEFT

Half a million downloads in a month. Today, our open source family takes another step forward. Thank you for the incredible support behind o…

I built a free, local Qwen-Image 2.1 app for Apple Silicon Macs: multi-image edits, transparent PNGs and a node canvas

ImbalancE: Inference-Time Latent Search Against Degree Imbalance in Link Prediction

InsightMap: Structured Spatial Modeling for Embodied Multimodal Reasoning

Knowing Is Not Choosing: What Explicit Verification Adds Beyond Generative Preference

LazySloth: Bounded LLM-based Lazy Tree Search for Fast Long Video Comprehension

Learning Continuous Patient Trajectories from Electronic Health Records

Multi-Class, Multi-Tier Network Intrusion Detection: A Comprehensive and Reproducible Benchmark

OTT3R: Multi-View 3D Reconstruction and Fast Dataset Generation at 1% Compute

Point2Part: Unified 3D Partitioning from Point Prompts

Qwen3.8 flash next ISTA-DASLab GGUF 50t/s TG and 1500t/s PP with 12GB VRAM and 64GB RAM Laptop on 'Strata' engine

Reducing the Adaptation Gap Through Reachable Fisher Geometry

Relevance Is Not Sufficient Evidence: Detecting Evidence Gaps Before Generation in RAG

Routing Should Pay for Itself: Sparse Supervision for Economical LLM Routing

SafeLLM4SE: Statistical Evaluation and Reporting for LLM-based Software Engineering Systems

Safer Content or Firmer Refusals? A Hybrid Perturbation Defense for Alignment under Harmful Fine-tuning

Video2Skill: From Streaming Experience to Reusable Embodied Skills

Visual Parallel Search: Learning to Search High-Resolution Images with Parallel Tile Inspection and Adaptive Zoom

Visual sensitivity is not claim retractability: persistence-aware credit assignment for multimodal reinforcement learning

vSkipper: Translating Dynamic Layer Skipping into LLM Serving Gains

Weights Read and Write Features: Scalable Parameter Decomposition Grounded in Activation Space

Which papyrus HTR is good enough? Character-error-rate tolerance of four papyrological tasks on Greek texts

You Only Reprogram Once: Rethinking Prolonged Training for Visual Reprogramming

Activation Flow: Manufacturing Activations for Steering

CoDrive: Cross-Vehicle World-Consistent Video Generation with Precise Trajectory Control for Cooperative Driving

Composable Decoding on the Probability Simplex: Theory and Implementation

ControlTrace: Recovering Control Fields for Hidden-Content Recognition

Detection of Adversarial Attacks on Super-Resolvers Using Spectral Features

DPAMixerSR: An Efficient Degradation-Pattern-Aware Model for Image Super-Resolution